Study Of Building Electrical Experimental Platform Fault Diagnosis Intelligent Technology

The Building Electrical System is an important part of the building and it is a key element to make the building play its function. Once the Building Electrical System malfunction, it will cause the important electrical equipment stop, fire and other bad consequences which will bring inconvenient to people’s work and life, even cause the economic loss and personal casualties. At present, the research on Building Electrical System fault diagnosis is still in the blank at both home and abroad. This is because with the development of information technology, the function of Building Electrical System also improve and has become more huge and complex, but its importance and complexity has not received enough attention. Nowadays, the Building Electrical System is very different from the traditional way, along with the electrical equipment increased dramatically; the old detection methods need to face challenges. According to the characteristics of Building Electical System, the new fault diagnosis technology should be developed.This paper selects Building Electrical System as the research object, to find the best intelligent fault diagnosis algorithm to Building Electrical System. According to the character of Building Electrical System, which is difficult to establish mathematical model and with many subsystems, choosing the neural network as intelligent diagnosis algorithm and choose the more applied method of RBF and BP neural network to do comparative study and find out the best solution. To test the two algorithm’s diagnosis effect, use the same fault data which is gathering from building electrical experimental platform. This platform is made to simulate the real residential model and have low voltage distribution systems in common buildings, through the artificial settings of electrical fault; it can effectively verify the diagnosis effectiveness of the algorithm. This study laid the foundation for the engineering application.In addition, in order to realize intelligent on-line monitoring and fast diagnosis evaluation of Building Electrical System, the data acquisition system which is based on the embedded ARM controller was set up and built embedded Linux system development environment, make out the application for embedded data acquisition. To solve the problem of multi-loop sensor placement, proposed a sensor placement method based on the graph theory, which can get the best number and location of sensor, so that the system can be observed.After the several rounds of experiments based on building electrical experiment platform and studying of fault data, it is can be found that RBF neural network with quick convergence speed and stability diagnosis effect is more suitable as a intelligent algorithm on the fault diagnosis for the building electrical experiment platform.